Virtual InfoCubes are used to retrieve and display data from non business information warehouse (BW) data sources. Virtual InfoCubes can process data without having to load the data into BW structures. Calculations may be applied to the result set before it is passed to an online analytical processor (OLAP).
Requests from BW are based on BW InfoObjects. There are two types of InfoObjects: dimension characteristics and key figures. Dimension characteristics and key figures are typically represented as a star schema composed of dimension tables and fact tables.
A virtual InfoCube's user-defined function module will compose a structured query language (SQL) query based on the list of characteristics, the list of key figures, and the list of selections with the appropriate joins, aggregates, and filters. Ideally, the user-defined function module produces an optimal SQL query that would return as few rows as necessary for the specific request. Because BW typically is only aware of dimension hierarchy leaf-node data, a query will require a less than optimal operation that returns many more rows than the user needs to satisfy a query containing one or more dimension hierarchies.